RNA viruses caused the three major pandemics of the 20th century, influenza, polio, and HIV. Despite the central roles of RNA in essentially all diseases, the RNA structures that can potentially be therapeutic targets are largely unknown. The ultimate goal of this research is to predict reliably the structures of RNA molecules from their sequences by using knowledge of the interactions directing RNA folding. The results should also lead to rational design of therapeutics. The ability to predict RNA structure also facilitates interpretation of sequences determined by genome projects and other sequencing efforts. The foundation for structure prediction will be advanced by studies of the thermodynamic and structural properties of oligonucleotides. Particular emphasis will be placed on the sequence dependence of stability and local three dimensional structures of internal loops. Secondary structure prediction by free energy minimization will be augmented with constraints from NMR. The NMR assisted prediction of secondary structure (NAPSS) method uses two unassigned spectra to restrict folding space by identifying helixes of canonical base pairs. This allows determination of pseudoknots, which are difficult to predict with other methods. Pseudoknots usually have significant functions, and are thus potential therapeutic targets. An approach is proposed that restricts folding space further on the basis of NMR chemical shifts. NAPSS is also a first step in assigning resonances for 3D structure determination. To advance predictions of 3D structure, benchmarks will be developed to test force fields and computational methods. For example, (1) the complete solution structure of a new and novel internal loop motif discovered in the previous grant period will be determined by NMR, (2) a less stable structure in equilibrium with this novel structure will also be determined, and (3) structural aspects of single stranded oligoribonucleotide tetramers will be determined by NMR and compared to predictions from molecular dynamics simulations. The power of predictive methods will be tested by experimentally determining the secondary and 3D structures of regions of influenza RNA predicted to fold into stable secondary structures. Initial results indicate there is a conformational switch between a pseudoknot and hairpin that may regulate splicing. Microarrays of short oligonucleotides and small molecules will be used to discover compounds that target the discovered structures and could potentially serve as therapeutics for natural strains and any engineered for bioterrorism. These approaches provide the foundation for rapidly designing therapeutics once a genome has been sequenced.